Report

Bit Operations • C is well suited to system programming because it contains operators that can manipulate data at the bit level – Example: The Internet requires that bits be manipulated to create addresses for subnets and supernets • Two categories of bit operators – Logical bitwise operators – Shift bitwise operators Bit operations A bit can have one of two values: 0 or 1. The C language provides four operators that can be used to perform bitwise operations on the individual bits of any integer data type (signed or unsigned char, short, int, and long): Bitwise and Bitwise inclusive or Bitwise exclusive or One’s complement & | ^ ~ & and | are binary operators Require integral operands: integer or char Bit by bit comparison between the two operands Bit operations and: The and (&) operator takes two operands. A bit in the result is 1 if and only if both of the corresponding bits in the operands are 1. For example: x = 01100010 y = 10111010 x&y = 00100010 or: The or (|) operator also takes two operands. A bit in the result is 1 if and only if either one or both of the corresponding bits in the operands are 1. For example x = 01100010 y = 10111010 x|y = 11111010 Bit operations Differences between && and & operators: Logical and (&&) • Result of && is always 0 or 1 • The evaluation of an expression that uses && terminates if the first operand is false (zero) Bitwise and (&) • The result of & can be any number • Expressions with the & operator are always completely evaluated Bit operations Differences between || and | operators: Logical or (||) • Result of || is always 0 or 1 • The evaluation of an expression that uses || terminates if the first operand is true (1) Bitwise or (|) • The result of | can be any number • Expressions with the | operator are always completely evaluated Bit operations xor: The xor (^) operator takes two operands. A bit in the result is 1 if and only if exactly one of the corresponding bits in the operand is 1. For example x = 01100010 y = 10111010 x^y = 11011000 Bit operations • The complement operator is unary in that it requires a single operand. The other three are binary and require two operands. complement: The operator complement (~) converts 0 bits to 1 and 1 bits to 0. For example, x = 01100010 ~x = 10011101 Bit operations In addition to the boolean operations, C also provides binary operators that can be used to shift all the bits of any integer type to the left or the right. The left shift operator (<<) is used to shift bits to the left. The left operand is the integer value to be shifted, and the right is the number of bits to shift it. For example, x = 01100010 x<<4 = 00100000 n A left shift of n bits is a very efficient way to multiply by 2 Bit operations The right shift operator (>> )is used to shift bits to the right. The operands are identical to those used in left shift. The left operand is the bit string and the right operand is the number of bits to shift. Bit operations The right shift operator (>> ) (cont’d) However, an important difference is that the value of new bits shifted in from the left is machine dependent. Nevertheless, it is commonly the case that the new bit will be 0 unless (1) the left operand is a signed integer and (2) the high order bit was originally a 1. In this case the integer value represents a negative number and the shifting in of 1 bits keeps it negative. An example of the right shift: x = 01100010 x>>4 = 00000110 n A right shift of n bits is a very efficient way to divide by 2 Bitwise Use • Used as flags: 0 is off, and 1 is on • Use bit mask to set and test flags • Mask – variable or constant, usually stored in a byte of short integer, that contains a bit configuration used to control the setting or testing of bits in a bitwise operation Example of a bit mask in a short integer 1000 0000 1000 0001 Bit 15 Bit 0 Creating Masks • One-bit mask examples: – mask5 = 1 << 5 00000001 << 5 -> 00100000 – mask1 = 1 << 1 00000001 << 1 -> 00000010 – mask = mask1 | mask5 00000010 | 00100000 -> 00100010 Masking Masking is a commonly used technique in processing individual bits of an integer word. In this technique, bits in the mask corresponding to bits of interest in the target word are set to 1 and the others are set to 0. The mask and the target word are then and’ed. Here are some examples of masking (assume x is an integer): if ((x & 0x1) == 1) printf("x is odd\n"); /* Simple test for even/odd */ right2 = x & 0x3; /* Extract right 2 bit values */ bit10 = (x>>10) & 0x1; /* Extract bit 10 value righthalf = x & 0xffff; */ /* Extract right half of integer*/ Creating Masks Suppose we want to turn off the 5 leftmost bits of a number stored in an 8-bit byte. The mask should start with the five zero bits to turn off the first 5 bits and then contain 3 one bits to leave the last three unchanged. Operator & number xxxxxxxx mask 00000111 ------------result 00000xxx Creating Masks Example: We know that in ASCII code, uppercase and lowercase letters differ by only 1 bit in bit 5. The lowercase letters have a 1 in bit 5, the uppercase letters have a 0. It is easy to change a lowercase letter by using an appropriate mask to turn off bit 5. char 'a' mask result 'A' 01100001 11011111 ------------01000001 Masks • The bitwise and operator (&) is used to force masked bits in a variable to 0. • The bitwise inclusive or operator (|) is used to force masked bits in a variable to 1. • The bitwise exclusive or operator (^) is used to force masked bits in a data item to change. Exercise: Find the mask that when used with the & operator sets the second and fourth bits of an 8-bit integer to zero. Exercise: Find the mask that when used with the | operator sets the second and fourth bits of an 8-bit number to 1. Nibble The term "nibble" is used to describe 4 bits. In particular, a byte (8-bits) is composed of 2 nibbles. A 4-byte (32 bit) integer is composed of 8 nibbles. So we might view an integer as follows: nibble 0 nibble 1 nibble 2 nibble 3 nibble 4 nibble 5 nibble 6 nibble 7 The decimal value "294610" in binary (base 2) is equal to 1011100000102. In memory the value would be stored as illustrated by the following 8 nibbles: 0000 0000 0000 0000 0000 1011 1000 0010 Note that nibble 7 has the value 00102, and nibble 5 has the value 10112. Conversions In addition to decimal (base 10) and binary (base 2), hexadecimal (base 16) is commonly used to represent values in our field. A single digit in hexadecimal can be represented in 4 bits (a nibble). The binary and hexadecimal values of the decimal values 0-15 are shown in the following table. Also included are the octal (base 8) values: Hexadecimal Note that the letters "A-F" are used to represent the 6 additional "digits" in the hexadecimal system. Keep in mind that 15 in decimal, and 1111 in binary and F in hexadecimal are ALL the same value -- they are just different ways of representing the same value (we could extend this to roman numerals, i.e. "XV" -- this is just another representation of the same number). In C we can represent a hexadecimal value by prefixing it with a "0x". For example, "x=0xf;" sets "x" to the decimal value 15 (or the binary value 1111). The statement "x=0xff" would set x to the binary value "11111111", which is 255 in decimal. Hexadecimal When we do masking, we often find that using the hexadecimal notation is useful (since it corresponds more directly to bits and nibbles). We can ask C to print a value in hexadecimal (as opposed to decimal) by using the "%X" format item, for example if y=12, then printf("y=%d (%X)\n", y, y); would print the value of y in both decimal and hexadecimal, e.g.: y=12 (C)